Richard Feynman Computer Heuristics Lecture (1985) [video] | Hacker News Reader

Richard Feynman Computer Heuristics Lecture (1985) [video]

youtube.com

142 pointsespeed13 comments

At 1:08:35 Feynman tries to put his glasses in his t-shirt, thinking he has on a dress shirt with a pocket. He plays it off by rubbing the glasses against his t-shirt. Pretty awesome how he's still making me smile and chuckle from the grave.

Hats off to you Mr. Feynman. Your output may have been finite, but its effect is limitless.

https://www.youtube.com/watch?v=EKWGGDXe5MA&t=1h8m35s

I like how he describes chess about 44 minutes in. Makes me wonder how modern chess engines like stockfish rate a certain board configuration. I expect it is quite more nuanced than his description of dead/not dead and summing up the material on both sides.

> I expect it is quite more nuanced than his description of dead/not dead and summing up the material on both sides.

It is a bit more nuanced. There are heuristics ( breadth and depth searches ) which assigns positional values and also opening and end game database searches. Using remaining material values is the most basic form of chess engine. If that's all stockfish did, most chess players would beat stockfish.

I built a very simple chess engine for my AI class. I started off with the basic "material values". Then added basic heuristics. Then added database lookups.

Now with neural networks and machine learning, chess engines are even more sophisticated.

There is a fascinating (and charming) paper by Alan Turing that describes his "Turochamp" chess 'engine'. Apparently, it was the first program capable of playing a complete game of chess, and the first program that could be described as a computer game (although it sadly only ever existed on paper). The general pattern he outlines (a heuristic evaluation function with hand-tuned weights, along with minimax game tree search--i.e. backwards induction) has formed the basis of most chess engines, both ancient and modern. Here's the original copy: https://docs.google.com/file/d/0B0xb4crOvCgTNmEtRXFBQUIxQWs/...

Intriguingly, Turing posed the question, "Could one make a machine to play chess, and to improve its play, game by game, profiting from its experience?" This reinforcement learning approach to chess did not enjoy much success--until AlphaZero. That story that has been well-told in many places, but perhaps best so by David Silver in this recently released lecture by DeepMind: https://www.youtube.com/watch?v=ld28AU7DDB4. The first ~40 mins are a lucid explanation of the classical methods, and the rest covers RL/MCTS/AlphaZero.

>..."Could one make a machine to play chess, and to improve its play, game by game, profiting from its experience?" This reinforcement learning approach to chess did not enjoy much success--until AlphaZero.

Don't forget Samuel's computer checkers program from 1959. It was among the world's first successful self-learning programs.

https://en.wikipedia.org/wiki/Arthur_Samuel

Interesting that he noted pattern recognition as the limitation for computers.

It would be really interesting to know what Feynman would make of the current state of machine learning. It's awesome what tasks can be performed with it, but I imagine he would be disappointed with our level of understanding of how these systems work.

Why would you say we don't understand how these systems work? Stochastic gradient descent, for example, is not particularly enigmatic. Pattern Recognition and Machine Learning by Christopher Bishop is a good place to start if you want to gain an understanding of how and why machine learning algorithms work.

He's talking about understanding their process, not their mechanics, machine learning systems are usually black boxes with no guarantees, we run into issues with this fact regularly because all we can do is train them and then study the results, it can't tell us anything for certain.

Because of emergent phenomena. That’s like saying you understand how organisms work because you understand how molecules work. Look at Wolfram’a New Kind of Science and even finite automata can have amazing patterns.

A small discussion from 2014: https://news.ycombinator.com/item?id=7457172

Note the way he picks on a woman in the audience for no reason to her surprise and bemusement. If you've read "Surely You're Joking..." you'll recognise it as one of his pick-up techniques. I'm a massive Feynmann fan, he has such insight, wisdom and humour. All of us have foibles and we see (IMHO) one of his in this one. They're all adults and I can completely forgive him for it just as I could if it was him putting his index finger up his left nostril and having a good dig while on camera. Not super pleasant to watch though. Adults deciding or not to have sex is fine. Being jerks to try and do it, eh, no more than yuk but, yeah, yuk.

If you're going to make a comment like that, about a single moment in a video over an hour long, you could try and be kind to the reader and at least say where in the video this occurs.

Reply on news.ycombinator.com